Pancreatic Ductal Adenocarcinoma (PDA) is a uniquely lethal disease for which there is no cure. Current drug regimens prolong survival by just weeks and even surgical removal of 'resectable' PDA fails to prevent disease recurrence. The discovery of new drugs for this or any cancer is hindered by a lack of truly predictive preclinical screening tools. Recent advances in genetic engineering have allowed the creation of sophisticated mouse models that recapitulate both the genotypic and phenotypic hallmarks of human cancers. I propose to use a recently-described mouse model of PDA to evaluate the ability of several agents to cause reversion of established PDAs. This model is based on the conditional mutation of K-ras and p53, two of the most frequently mutated genes in pancreatic cancer. Effects on tumorigenesis will be tracked via several imaging modalities currently under investigation. In order to validate this model, the PDA mice will initially be treated with drugs previously tested in human patients. Following this, novel agents targeting relevant pathways will be evaluated in a high-throughput setting. [unreadable] [unreadable] [unreadable] [unreadable]